The present invention generally relates to the discovering of failures in complex systems for data processing purposes.
Discovering of failures in complex systems, i.e. systems with a plurality of subsystems, generally exhibits difficulties since there might be a plurality of difficult failures in the various subsystems, and the failures might also reveal in different forms. Failures which can obviously be seen, such as when a system is completely inoperable, can easily be detected by the user of the system. However, in case that internal parameters within the system or the subsystems change and thus influence the entire system, this might not be detected from outside, e.g., by the user. These failures might only turn out by their effects and can solely be detected by trained users, special instructed persons, or by sp cific inspection routines.
Some failures occurring in one subsystem only have an impact on the subsystem itself e.g. in a way that only the subsystem might become inoperable. However, other failures occurring in one subsystem might also influence other s bsystems and thus the entire system.
Examples for systems which are in particular sensitive for failures are automated test equipment (ATE) systems, such as IC testers, or sorting and packaging machines with optical recognition of products, or the like. Defects or failures in one subsystem can influence the entire ATE system in a way that the testing results can be erroneous and might not represent the actual state of a device under test (DUT). This can lead to a reduced yield of the DUTs to be tested and components which are well functioning might be selected as defective, and vice versa.
Failures in complex systems are usually detected in the art by specific diagnosis routines monitoring the functional ability (functionability) of the system. The diagnsis routines are normally started in defined service or maintenance intervals, or before use of the system. The execution of such a diagnosis routine is to be started by the user or might be automatically and periodically started by the system. The effective using time of the system is reduced by the time required for executing the diagnosis routines. In case the system is used for production or testing purposes, the time required for executing the diagnosis routines might increase the product costs of the devices to be produced or tested by the system. Furthermore, failures occurring during the operation of the system might not be recognized and can lead to a series of faulty products.